Model: SoLU Model: 4 Layers, 2048 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 425 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.7162. Min Act: -0.0006

Data Index: 1230249 (C4 (Web Text))

Max Activating Token Index: 781

Text #1

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6100. Min Act: -0.0006

Data Index: 975185 (C4 (Web Text))

Max Activating Token Index: 885

Text #2

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6389. Min Act: -0.0006

Data Index: 51305 (C4 (Web Text))

Max Activating Token Index: 752

Text #3

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6452. Min Act: -0.0006

Data Index: 68142 (C4 (Web Text))

Max Activating Token Index: 760

Text #4

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.5845. Min Act: -0.0006

Data Index: 1014787 (C4 (Web Text))

Max Activating Token Index: 997

Text #5

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6040. Min Act: -0.0006

Data Index: 58676 (C4 (Web Text))

Max Activating Token Index: 816

Text #6

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6339. Min Act: -0.0006

Data Index: 205738 (C4 (Web Text))

Max Activating Token Index: 658

Text #7

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6692. Min Act: -0.0006

Data Index: 79206 (C4 (Web Text))

Max Activating Token Index: 845

Text #8

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6233. Min Act: -0.0006

Data Index: 500706 (C4 (Web Text))

Max Activating Token Index: 210

Text #9

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6040. Min Act: -0.0006

Data Index: 684510 (C4 (Web Text))

Max Activating Token Index: 90

Text #10

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6018. Min Act: -0.0006

Data Index: 1186264 (C4 (Web Text))

Max Activating Token Index: 354

Text #11

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6092. Min Act: -0.0006

Data Index: 1081568 (C4 (Web Text))

Max Activating Token Index: 665

Text #12

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.6059. Min Act: -0.0006

Data Index: 1009782 (C4 (Web Text))

Max Activating Token Index: 724

Text #13

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.5428. Min Act: -0.0006

Data Index: 825137 (C4 (Web Text))

Max Activating Token Index: 51

Text #14

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.5602. Min Act: -0.0006

Data Index: 953185 (C4 (Web Text))

Max Activating Token Index: 460

Text #15

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.4655. Min Act: -0.0006

Data Index: 659319 (C4 (Web Text))

Max Activating Token Index: 920

Text #16

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.5668. Min Act: -0.0006

Data Index: 1087345 (C4 (Web Text))

Max Activating Token Index: 459

Text #17

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.5064. Min Act: -0.0006

Data Index: 1144172 (C4 (Web Text))

Max Activating Token Index: 974

Text #18

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.5385. Min Act: -0.0006

Data Index: 1142625 (C4 (Web Text))

Max Activating Token Index: 589

Text #19

Max Range: 0.7162. Min Range: -0.7162

Max Act: 0.5521. Min Act: -0.0006

Data Index: 1003249 (C4 (Web Text))

Max Activating Token Index: 241